fault tolerance algo superposion

Question

I have seen many options where CEPH works with docker container or docker swarm.But in my use case, I require to run many instances of ceph to learn about the fault tolerance algorithm called paxos that have many variantes.

According to the documentation of docker swarm , it says that docker uses a concensus algo called raft;

will both algorithms have an impact one on the other or is there a better solution other then VMs that would allow me to  run CEPH without having this interference in the distributed algorithm behavior. especially for the fault tolerance with the algo of CEPH called "Crash"

Answer 1

Hi@emna,

I don't know to much related to your use case. But I think this task can be done by Kubernetes or Openshift cluster very easily. It will allow to monitor the containers as well.